gemfreq type/token counts ratios

[Cynthia Audisio]

Hello Chibolts,

I've got a group of files, each of them's got several "gems" with play situations. Is it possible to get separate type/token totals and ratios for each of the gems in a file ?

This is how the file looks:

.

.

@Bg:    play1

.

.

.

@Eg:    play1

.

.

.

.

@Bg:    play2

.

.

.

@Eg:    play2

.

.

and what I need is individual type/token counts and ratio for each play situation (play 1, play 2, etc). Up to now I've run gemfreq which yields a freq list (not total number of type/token and type/token ratio, which is what i need).

Thanks,

[Leonid Spektor]

Cynthia,

First I need to get more information from you. Do your data files have an @ID: headers? Do you want to get type/token and type/token ratio for speaker words or for morphological analysis words or for lemmas? Do you want the output in plain readable text format or in Excel format? Different answers to those questions will require different commands to get the exact result that you want.

Please allow me to explain the reason for my second question. For example, if you have the following sentence:

*MOT: you can't put it on the table and table it.

If you run FREQ on speaker words, then you will get result:

  1 and

  1 can't

  2 it

  1 on

  1 put

  2 table

  1 the

  1 you

------------------------------

    8  Total number of different item types used

   10  Total number of items (tokens)

0.800  Type/Token ratio

If you run FREQ on morphological analysis words, then you will get result:

  1 coord|and
  1 det:art|the
  1 mod|can
  1 neg|not
  1 n|table
  1 prep|on
  2 pro:per|it
  1 pro:per|you
  1 v|put&ZERO
  1 v|table
------------------------------
   10  Total number of different item types used
   11  Total number of items (tokens)
0.909  Type/Token ratio

If you run FREQ on lemmas, then you will get result:

  1 and
  1 can
  2 it
  1 not
  1 on
  1 put
  2 table
  1 the
  1 you
------------------------------
    9  Total number of different item types used
   11  Total number of items (tokens)
0.818  Type/Token ratio

Leonid.

--
You received this message because you are subscribed to the Google Groups "chibolts" group.
To unsubscribe from this group and stop receiving emails from it, send an email to chibolts+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/65a859f6-26e9-487f-9246-761b46de9bb3o%40googlegroups.com.

[Cynthia Audisio]

Hello Leonid,

Our data have @ID: headers and we want to get type/token totals and ratio for *speaker words*. It would be of great help to get the output as a spreadsheet. Gemfreq does almost all the work. The only missing information is the one in blue in the following sample output:

From file <........>

  3 tiers in gem "disputa_1":

      1 abuela

      2 dije

      1 duele

      1 eso

      1 heladito

      1 idea

      2 la

      2 mamá

      1 me

------------------------------
    9  Total number of different item types used
   11  Total number of items (tokens)
0.818  Type/Token ratio

From file <........>

  4 tiers in gem "disputa_2":

      3 no

      1 porque

      1 sí

      2 tenía

      1 un

      1 yo

------------------------------
    9  Total number of different item types used
   11  Total number of items (tokens)
0.818  Type/Token ratio

etc.

To unsubscribe from this group and stop receiving emails from it, send an email to chib...@googlegroups.com.

[Leonid Spektor]

Cynthia,

I would suggets the following commands to extract the right gems from your data files:

gem +d +t@ID +sdisputa_1 +fdisputa_1 filenames.cha

gem +d +t@ID +sdisputa_2 +fdisputa_2 filenames.cha

 etc...

Next run freq +d3 *.disputa*.cex on resulting GEM output files. The Excel output will be called stat.frq.xls.

Leonid.

To unsubscribe from this group and stop receiving emails from it, send an email to chibolts+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/7834c54f-e966-4f88-95db-2d76d2e0e743o%40googlegroups.com.

[Cynthia]

I'll try that. Thank you!

You received this message because you are subscribed to a topic in the Google Groups "chibolts" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/chibolts/2r8Ifpxd-44/unsubscribe.
To unsubscribe from this group and all its topics, send an email to chibolts+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/chibolts/239D98C6-A98A-46BF-A7AE-A0C68A9841C8%40andrew.cmu.edu.